►
From YouTube: Keptn Community Meeting - December 9th, 2019
Description
Keptn 0.6.0.beta2
https://github.com/keptn/keptn/releases/tag/0.6.0.beta2
Keptn specification 0.1.1
https://github.com/keptn/spec/releases/tag/0.1.1
KEPs
https://github.com/keptn/enhancement-proposals
Release plan for 0.6.0
https://github.com/keptn/keptn/milestone/11
A
So,
let's
get
started
with
the
release
we
did
last
week.
It
contains
well
small
enhancements
and
a
couple
of
bucks
fixes
for
them
from
the
previous
zero
six
better.
So
on
the
one
hand,
it
implements
the
captain
specification,
0
1
1,
and
we
will
dive
into
the
details
about
that
in
a
couple
of
seconds.
We
have
some
new
features,
for
example,
that
we
now
also
support
nginx
on
EPS,
that
we
allow
you
to
configure
a
custom
domain
when
using
nginx,
and
not
only
when
using
sto
the
lighthouse
service
forwards,
the
custom
data
to
SLI
provider.
A
So
this
is
a
very
important
feature
for
all
the
third-party
system:
integrators
that
want
to
or
have
to
actually
forward
the
data
that
is
necessary
in
the
SLI
provider.
So
my
best
example
is
from
from
our
good
friend
Hendrik
from
the
Neo
load
service.
So
neo
load,
for
example,
has
a
unique
test
ID
with
each
test
run
and
if
you
want
to
get
a
matrix
for
that
specific
test
run,
you
need
to
have
that
test
ID
in
the
SLI
provider.
So
we
now
provided
a
functionality
that
you
can
add.
A
Custom
enter
Henrik,
so
we
we
can
provide
custom
data
in
the
index
Li
in
this
in
the
events
and
that
is
then
propagated
to
the
SLI
provider.
So
you
have
to
date
the
way
you
need
it
and
we
also
extended
the
event
so
that
the
deployment
is
included
through
the
SLI
providers.
That
was
needed,
because
if
you
do
a
Bluegreen
deployment
you
need
to
specify
for
which
version
you
would
actually
like
to
gather
as
allies.
So
in
this
case
you
can
select.
A
Do
you
want
to
gather
SLI
metrics
for
the
canary
version
of
the
primary
version?
We
got
rid
of
the
fluent
bit
piece
because
we
were
not
actually
using
it.
Then
it
was
just
like
a
I
think
that
is
still
installed
every
time
we
set
up
captain,
but
no
one
uses
it
and
it
it
will
help
us
I'm
slim
down
the
resource
footprint.
Again
there
is
a
chat
message.
A
A
A
A
So
previously
this
was
like
just
the
the
timeframe
so
like
for
the
last
10
minutes,
and
now
you
can
specify
start
and
end
date,
time,
which
will
most
likely
also
make
it
way
easier
to
integrate
into
existing,
let's
say,
CD
tools
so,
for
example,
in
implementing
the
quality
gates
use
case
in
a
Jenkins
pipeline,
for
example,
where
you
of
course
have
to
start
and
end
date.
Time
of
your
tests.
A
Anyways,
we
fixed
the
issue
with
node
ports
that
are
generated
in
the
health
chart
and
there
are
two
known
limitations
as
of
now
so
quality
gates
standard
on
use
case
right
now
doesn't
work
on
OpenShift
and
if
I
remember
correctly,
there
is
something
to
do
with
with
the
ingress
configuration
on
OpenShift.
We
I
think
have
it
in
the
release
4:06
in
there
that
we
will
use
the
OSI
routes
for
that.
But
there
is
some
research
necessary
and
in
fact
this
is
like
only
a
syntax
error.
A
So
we
we
always
talk
about
a
new
artifact
event,
also
in
the
CLI,
where,
in
fact,
if
you
say
sin,
even
new
artifact,
it
actually
sends
a
configuration
change
event.
So
we
will
we
will.
We
will
find
a
way
to
to
get
rid
of
that
small
glitch,
all
right
so
much
to
to
the
0-6
better
to
release.
There
is
another
question:
can
I
will
in
future
be
able
to
adjust
the
deployment
and
testing
strategy
once
a
project
and
services
are
already
on
board?
It.
A
B
Actually,
this
is
untested.
For
example,
you
consider
you
have
a
direct
deployment
now
and
then
changing
to
blue
green
deployment.
We
have
to
generate
some
additional
resources
like
virtual
services,
destination
routes
and
so
on,
and
what
we
have
to
do
is
to
verify
better.
This
resources
are
also
correctly
generated,
men
which
changed
deployment,
information
during
after
services
installed
or
onboard.
So
we
still
need
a
issue
which
says
to
verify
this
technically,
it
should
be
possible.
A
A
These
two
things
have
changed
and
we
already
have
that
in
the
in
the
better
release.
Also,
so
we
have
now
a
deployment
field.
Forget
SL,
I,
honestly
don
t
to
distinguish
between
the
primary
and
the
cannery
version
when
retrieving
SL
eyes
for
a
service
and
we've
added
a
description
for
the
remediation
strategy
in
the
shipyard
file.
A
So?
Can
I
I
really
like
this
discussion
about
Canaries
and
Bluegreen
deployment,
so
so,
basically,
cannery
for
me
is
a
release
strategy
where
your
shift
traffic
to
to
a
to
a
second
deployment
that,
in
our
case,
is
now
called
cannery.
So
this
could
be
somewhat
confusing.
I
understand
that
in
fact,
we
haven't
implemented
cannery
releases,
yet
we
will
wait
until
we
have
the
workflow
engine
for
that,
because
otherwise
it's
it
would
be
too
too
much
logic
and
too
complicated
in
the
circuit.
That
does
that
about
the
cannery
deployment.
A
This
is,
could
you
could
compare
it
to
a
Bluegreen
deployment
basically,
and
we
support
that
already,
it's
the
remote
mock-up
for
the
new
series,
six
captain
bridge
planned
for
January.
Yes,
there
is-
and
in
fact
I
can
pull
this
up
later
on.
It's
a
good
point.
So,
if
I
forget,
please
remain
remind
me
and
write
this
word
that
changes
in
the
specification.
A
Now
we
have
a
new
a
new
repository,
it's
called
captain
enhancement
proposals,
and
this
is
this
process.
So
we've
we've
looked
around
at
other
open
source
projects
and
how
they
handle
their
specification
and
changes
in
the
specification,
and
we
looked
in
detail
on
the
open,
telemetry
project
and
they
have
also
a
specification
project
very
handle
the
specification
and
to
heaven
an
enhancement
proposals
project,
and
we
also
now
follow
that
approach.
A
So
if
you
want
to
have
a
change
in
the
specification,
we
would
ask
you
to
fill
out
an
enhancement
proposal
and
actually
it
it
tells
you
in
the
documentation
what
you
need
to
do.
So
we
call
them
caps,
captain
enhancement
proposals,
and
you
should
basically
fork
this
repository
and
write
down
your
proposal
in
the
text
directory
and.
A
Here
is
a
description
on
what
should
actually
go
into
a
cap
into
an
enhancement
proposal
and
what
should
actually
be
handled
in
an
issues
or
just
for
bug.
Fixes
issues
are
just
fine
if
you
want
to
have
some
behavioral
changes
of
captain
or
any
core
service.
This
is
something
that
would
need
to
go
in
a
captain
enhancement
proposals
of
our
example.
A
The
self-healing
use
case
was
a
captain
enhancement
proposal,
and
here
it
says
in
detail
what
you
need
to
do
to
write
one,
and
it
has
several
stages,
so
it
can
be
proposed,
approved
or
rejected,
and
here
it
says
we
we
borrowed
that
process
from
the
open,
telemetry
enhancement
proposal,
and
we
are
looking
forward
of
receiving
plenty
of
enhancement.
Proposals
from
your
site
so
feel
free
to
add
one
or
two.
A
A
So
we
want
to
have
an
investigation
of
helm
3,
since
it
was
released
a
couple
of
weeks
ago,
if
we
can
replace
it
with
the
current
helm
version
we
have,
which
would
also
make
the
installation
easier,
because
we
we
can't
get
rid
of
tiller,
but
this
is
more,
as
the
label
says,
a
research
story.
Now
we
want
to
improve
the
captain
installer
and
improve
the
waiting
times
for
ingress.
A
We
will
do
an
update
on
the
on
the
seal
I
so
that
you
can't
install
the
wrong
the
wrong
version
anymore.
So
we
had
problems
with
that
with
the
serial
six
better
release
with
the
first
one
that
the
Installer
all
of
a
sudden
didn't
install
the
latest
stable
version,
but
the
better
version.
So
we
need
to
to
improve.
On
that
end.
A
A
This
is
what
we're
going
to
do.
Then
we
will
improve
the
dynaTrace
service
so
right
now
the
dynaTrace
monitoring
setup
is
done
through
shell
scripts.
We
will
implement
that
through
the
CLI
as
much
as
possible
and
also
allow
to
set
up
management
zones
and
dashboards
for
services,
and
almost
all
of
the
the
auto
tagging
rules
automatically.
A
We
have
a
story
that
will
allow
you
to
install
a
kept
version
without
the
dependency
on
kubernetes.
So
from
what
I've
heard,
we
will
use
micro
Cades,
but
that
so
you
can
run.
Captain
is
a
single
process
on
a
VM
or
on
a
on
a
physical
host.
If
you
want-
and
you
don't
have
the
dependency
to
kubernetes
anymore,
because
that
is
abstracted
away
by
micro
Cades,
so
we
need
that
for,
for
example,
for
quality
gate
stand-alone
use
cases
where
there
is
no
kubernetes
anywhere
a
lot
of
project
names
with.
A
So
we
have
that
feedback
from
from
from
christian,
and
this
is
a
minor,
a
minor
issue.
We
have
in
one
of
our
services
and
we
will
get
rid
of
that.
It's
called
feature
here.
I
would
argue
it's
a
bug.
We
will
provide
a
functionality
to
expose
the
captain
bridge
so
right
now,
you
would
need
to
issue
the
cue
CTL
port
for
what
command
to
expose
the
bridge.
A
A
A
This
is
also
very
interesting
because
it
ties
together
with
the
issue
that
we
have
before.
So
we
will
make
sure
that
you
only
can
install
the
server
version
of
the
CLI
version
that
you
have
right
now
and
to
that,
and
we
also
want
to
have
the
ability
to
check
if
there
is
a
newer
version
of
captain
already
available
for
you,
and
we
will
do
that
through
the
CLI.
A
So
most
likely-
and
it
already
says
it
here-
you
can
opt
out,
but
once
a
day
we
would
check
if
there
is
a
new
version
of
the
captain
CLI
available
for
you.
That
brings
you
new
features,
update
to
helm
version
3,
so
to
sustain
the
implementation
story
to
the
to
the
research
task.
It's
not
yet
100%
sure
that
we're
going
to
do
that.
A
Simplify
release:
this
is
an
internal
issue.
We
will.
This
is
something
we
really
need
to
do
so.
The
college,
a
standalone
use
case-
is
not
documented
yet
so
you
will
find
the
documentation
for
that
one
in
the
tutorial
section,
and
there
are
two
packs
that
we
discovered
in
the
course
of
of
the
last
days
that
the
Nets
cluster,
when
the
Nets
cluster
is,
is
restarted
within
the
server
and
the
distributors
need
to
reconnect
to
the
server.
A
So
this
is
a
lot
of
work
that
we
have
planned
for
for
release
by
January,
twenty
twenty
twenty,
it's
a
nice
date
by
the
way
three
times
twenty
was,
and
that
is
basically
it.
So
are
there
any
questions
right
now
that
we
could
answer?
Who
there
is
someone
in
the
chat
just
enhancements,
meaning
the
enhancements
issues.
A
A
Because
if
not
I
actually
looked
up
the
calendar
when
the
next
community
meeting
is
going
to
be
and
December
23rd
is
when
most
of
our
guys
are
already
on
Christmas
holiday,
the
next
would
be
January
6th
next
year,
which
is
public
holiday.
That
leaves
us
with
January
20-20-20.
Witches,
coincidentally,
are
now
realized
that
the
date
of
the
release
of
the
planned
release
I
will
start
back
with
the
team.
A
If
we,
if
we
start
the
next
year
on
January
13th
or
if
we
stick
to
the
schedule
and
start
on
January
20,
there
is
one
more
message
in
the
chat
raba
I
will.
I
will
reach
out
to
you
afterwards
to
discuss
the
the
new
bridge.
Yes,
let
me
reach
out
to
you
afterwards
with
regards
to
the
to
the
river
questions.
A
So
this
is
how
it
will
look
like
and
if
you
want
take
a
screenshot
and
write
feedback
to
us
so
most
likely
this
pane
on
the
left
side
will
go
away
because
there
is.
We
had
a
lot
of
filters
in
there
initially,
but
we
don't
need
them
anymore.
So
up
here
you
will
have
a
captain
project
selector,
where
you
can
switch
between
different
projects.
A
Then
here
you
will
find
a
list
of
all
of
the
services
that
run
in
that
project
that
are
on
boarded.
You
get
an
overview
of
all
the
stages
of
the
overall
services
and
how
much
are
in
a
failing
state,
for
example,
and
the
rest
relevant
event
and
for
each
of
the
services
you
see
all
of
the
flows
either
if
it's
a
deployment
flow
or
problem
flow,
and
if
you
select
one
of
those
flows,
you
get
the
the
view
that
you
already
know.
So
you
get
all
of
the
events
that
belong
to
that
flow.
A
On
the
right
side,
you
have
a
mini-map
here
where
you
can
quickly
jump
to
the
stage
that
is
most
interesting
to
you
or
to
events
that
raced
an
error
or
other
important
events.
And
basically
you
can
then
go
into
detail
and
see.
What's
what
it
is
all
about
and
for
the
validation
for
example,
we
we
will
list
all
of
the
SL
ice.
If
you
want
to
expand
this
section,
and
you
will
also
have
a
visual
representation
of
the
of
the
how
the
SLI
performed
over
time.